Leukemia (2005) 19, 268–274 & 2005 Nature Publishing Group All rights reserved 0887-6924/05 $30.00 www.nature.com/leu
Th1-biased humoral immune responses against Wilms tumor gene WT1 product in the patients with hematopoietic malignancies F Wu1, Y Oka1, A Tsuboi2, OA Elisseeva2, K Ogata3, H Nakajima4, F Fujiki4, T Masuda1, M Murakami1, S Yoshihara1, K Ikegame2, N Hosen1, M Kawakami2, M Nakagawa5, T Kubota6, T Soma7, T Yamagami7, M Tsukaguchi8, H Ogawa1, Y Oji4, T Hamaoka9, I Kawase1 and H Sugiyama4 1 Department of Molecular Medicine, Osaka University Graduate School of Medicine, Osaka, Japan; 2Department of Cancer Immunotherapy, Osaka University Graduate School of Medicine, Osaka, Japan; 3Third Department of Internal Medicine, Nippon Medical School, Tokyo, Japan; 4Department of Functional Diagnostic Science, Osaka University Graduate School of Medicine, Osaka, Japan; 5Nissay Hospital, Nippon Life Saiseikai Foundation, Osaka, Japan; 6Tondabayashi Hospital, Osaka, Japan; 7 Department of Internal Medicine, Osaka Minami National Hospital, Osaka, Japan; 8Sakai Municipal Hospital, Osaka, Japan; and 9 Department of Oncology, Osaka University Graduate School of Medicine, Osaka, Japan
The Wilms’ tumor gene WT1 is highly expressed in leukemias and myelodysplastic syndrome (MDS), and WT1 expression levels increase along with the disease progression in chronic myeloid leukemia and MDS. We previously reported that IgM and IgG WT1 antibodies were detected with significantly higher detection rate and antibody titers in leukemias and MDS compared to those in healthy volunteers. In this study, whether IgG humoral immune responses against WT1 protein were Th1or Th2-type were determined by measurement of four subclasses of IgG WT1 antibody, IgG1, IgG2, IgG3, and IgG4. In leukemias and MDS, Th1-type WT1 antibodies such as IgG1, IgG2, and IgG3 were significantly increased in both detection rate and antibody titers compared to those in healthy volunteers, whereas Th2-type WT1 antibody such as IgG4 did not increase. These results showed that Th1-biased humoral immune responses against WT1 protein were generated in leukemias and MDS. These results should allow us to consider that Th1-biased cellular immune responses against WT1 protein, which was essentially needed for cancer immunotherapy targeting WT1, should be elicited in patients with hematopoietic malignancies. Leukemia (2005) 19, 268–274. doi:10.1038/sj.leu.2403539 Published online 11 November 2004 Keywords: Wilms’ tumor gene (WT1); humoral immune response; IgG subclass; Th1; cancer immunotherapy
Introduction The Wilms’ tumor gene WT1, first isolated as a gene responsible for a pediatric renal neoplasm, Wilms’ tumor,1,2 encodes a zincfinger transcription factor involved in cell proliferation and differentiation, in apoptosis, and in organ development. Although the WT1 gene was originally categorized as tumorsuppressor gene,3,4 we proposed that the wild-type WT1 gene performed an oncogenic rather than a tumor-suppressor function5 on the basis of the accumulated evidence such as (i) high expression levels of the wild-type WT1 gene in hematopoietic malignancies including leukemia and myelodysplastic syndrome (MDS) and in various kinds of solid tumors,6–9 (ii) growth inhibition of human leukemic and solid tumor cells by WT1 antisense oligodeoxynucleotides,8,10 and (iii) growth promotion and differentiation inhibition of murine myeloid progenitor cells by the constitutive expression of the WT1 gene.11,12 Correspondence: Professor H Sugiyama, Department of Functional Diagnostic Science, Osaka University Graduate School of Medicine, 1-7, Yamada-Oka, Suita City, Osaka 565-0871, Japan; Fax: þ 81 6 6879 2593; E-mail:
[email protected] Received 17 February 2004; accepted 1 September 2004; Published online 11 November 2004
The above findings indicated that WT1 product could become a target for cancer immunotherapy. In fact, we and others generated human WT1-specific cytotoxic T lymphocytes (CTLs) in vitro.5,13–16 Furthermore, it was shown that mice immunized with WT1 peptides or WT1 cDNA elicited WT1-specific CTLs and rejected challenges of WT1-expressing tumors,17,18 indicating that the WT1 protein could in vivo serve as a tumor rejection antigen. It was also shown that WT1-specific CTL precursors (CTLp) were detected more frequently in the patients with hematopoietic malignancies than in healthy volunteers.19 Recently, we and others detected IgM and IgG WT1 antibodies in the patients with hematopoietic malignancies.20,21 Disease progression of MDS from refractory anemia to overt leukemia was associated with isotype class-switching of WT1 antibody from IgM to IgG.21 These findings indicated that WT1specific cellular immune responses that induced isotype classswitching of WT1 antibody had been elicited, suggesting that WT1 protein was immunogenic. Thus these results should provide us with the rationale for elicitation of CTL responses against WT1 product in cancer immunotherapy targeting it.21 Determination of IgG subclasses of the antibodies against WT1 protein should lead to further understanding of the immune responses against WT1 protein in the WT1-expressing tumor-bearing patients, because IgG subclasses in humoral immune responses were associated with different helper T cells, Th1 or Th2, which stimulated B cells in different ways, respectively.22 In this study, IgG subclasses of WT1 antibody were analyzed to determine whether humoral immune responses against WT1 protein in patients with hematopoietic malignancies are Th1- or Th2-type.
Materials and methods
Patients Sera and peripheral blood mononuclear cells (PBMCs) were obtained from 96 patients with hematopoietic malignancies (28 acute myeloid leukemia (AML), 12 acute lymphoid leukemia (ALL), 16 chronic myeloid leukemia (CML) (eight in chronic phase (CP), eight in blast crisis (BC)) and 40 MDS) (Supplementary Table 1). The MDS patients included 15 with refractory anemia (RA), 15 with RA with excess of blasts (RAEB), and 10 with RAEB in transformation (RAEB-t). The patients had not received chemotherapy at the time point when blood samples were obtained, except the patients with CML in BC. The control group was comprised of 53 healthy volunteers who gave their informed consent for this study.
Th1-biased immune responses against WT1 F Wu et al
Reverse transcription-polymerase chain reaction for quantitation of WT1 expression levels RNA was prepared from PBMCs at diagnosis and converted into cDNA. Polymerase chain reaction (PCR) was performed for optimized cycles with a DNA thermal cycle as described previously.7,23 WT1 expression levels in the samples were shown relatively to that (defined as 1.0) in K562 leukemic cells.6,7,23
269 correlation between the presence of IgG subclasses of WT1 antibody and either sex or clinical performance of the patients. Linear regression coefficient (r) was used for evaluation of the correlation between WT1 antibody concentrations and either WT1 expression levels or patient ages. Statistical analysis was performed by JMP (SAS Institute, Cary, NC, USA) and StatView software.
Results
Dot-blot assay for quantitation of IgG subclasses of WT1 antibody Recombinant human WT1 protein hWT3 (1–294 amino acids) were produced and purified as described previously.21.Serum #223 obtained from an AML patient was used as a reference serum because it contained high titers of all the four subclasses of IgG WT1 antibody in preliminary experiments. Standardization of titers of IgG subclasses of WT1 antibody in the reference serum #223 was performed by calibration with heterologous interpolation as described previously.21,24–26 Briefly, the mixture of mouse anti-human kappa and lambda antibodies (BDBiosciences Pharmingen, San Diego, CA, USA) was bound on nitrocellulose membrane Optitran (Schleicher & Schuell, Dassel, Germany), and WT1 protein (hWT3) was bound to another membrane. After blocking with 4% skim milk, the two membranes were loaded onto dot-blot apparatus (Schleicher & Schuell, Dassel, Germany). Two-fold serially diluted purified human myeloma proteins (IgG1, IgG2, IgG3, and IgG4) (Athens Research and Techenology, Inc., Athens, GA, USA) were applied to the membrane coated with the mixture of mouse anti-human kappa and lambda antibodies. Two-fold serial dilutions of the reference serum #223 were applied to the other membrane coated with hWT3. After washing, the two membranes were reacted with horseradish-peroxidase-conjugated sheep anti-human IgG1, IgG2, IgG3, and IgG4 (The Binding Site Limited, Birmingham, UK), followed by incubation with the substrate solution, Renaissance (NEN life Science Products, Boston, MA, USA), and exposure to Hyper film (Amersham Pharmacia Biotech, Buckinghamshire, England) with the same exposure time. Densities of dot blots were measured in densitometric units with a computerized scanning analyzer system (Molecular Dynamics, Sunnyvale, CA, USA). Concentrations of IgG1, IgG2, IgG3, and IgG4 WT1 antibodies in the reference serum #223 were determined as 28.7, 87.2, 3.7, and 2.5 mg/ml, respectively, from the respective calibration curves of purified human IgG1, IgG2, IgG3, and IgG4 myeloma proteins using Abelbeck software. Concentrations of IgG subclasses of WT1 antibody of each serum sample were determined according to the calibration curves for IgG subclasses of WT1 antibody of reference serum #223, as described previously,25,27 using Abelbeck software.
Statistics Cutoff values of individual IgG subclasses of WT1 antibody were determined to be 10, 40, 1.5, and 1.5 mg/ml for IgG1, IgG2, IgG3, and IgG4, respectively, on the basis of two receiveroperating characteristic plots.28 Mann–Whitney test was used for the measurement of statistical difference in WT1 antibody titers between two groups. Fisher’s exact test was used for the measurement of statistical difference in detection rate of WT1 antibody between two groups and for evaluation of the
WT1 expression levels in PBMC of patients with hematopoietic malignancies To demonstrate that WT1 expression levels, which reflected the amount of leukemic tumor burden, increased in company with the disease progression of CML and MDS, indicating an increase in an antigenic stimulation of immune system by WT1expressing cells along with the disease progression, WT1 expression levels were measured by quantitative reverse transcription (RT)-PCR in 72 of 96 patients (Supplementary Figure 1). In 70 of 72 patients with hematopoietic malignancies (25 AML, 10 ALL, 16 CML, and 19 MDS) except for two RA patients, WT1 expression levels were X1 105, while they were o1 105 in all of 53 healthy volunteers. WT1 expression levels increased in parallel with disease progression from CP to BC in CML, and from RA to RAEB and further to RAEB-t in MDS. These results confirmed previous reports of ours.6,7,23
Detection of IgG subclasses of WT1 antibody in sera from patients with hematopoietic malignancies IgG subclasses of WT1 antibodies were measured by dot-blot assay using reference serum #223 in 96 patients with hematopoietic malignancies (28 AML, 12 ALL, 16 CML, and 40 MDS) and 53 healthy volunteers (Table 1, Figure 1). IgG1 WT1 antibody was detected in 35 (36.5%) of the 96 patients with hematopoietic malignancies, whereas they were detected in only four (7.5%) of the 53 healthy volunteers (Table 1, Figure 1a). Both detection rate (Po0.0001) and titers (Po0.0001) of IgG1 WT1 antibody were significantly higher in patients with hematopoietic malignancies than in healthy volunteers. When detection rates of IgG1 antibody in individual disease of hematopoietic malignancies were compared with those in healthy volunteers, they were significantly higher in patients with AML (Po0.001), CML (Po0.01), or MDS (Po0.001) than in healthy volunteers (Figure 1a). The antibody titers of IgG1 WT1 antibody were also significantly higher in the patients with hematopoietic malignancies except ALL than in healthy volunteers (Figure 1a). Both detection rate (Po0.05) and titers (Po0.05) of IgG2 WT1 antibody were significantly higher in the patients than in healthy volunteers (Figure 1b). When detection rate and antibody titers of IgG2 WT1 antibody in individual disease were compared with those in healthy volunteers, both (Po0.05 for detection rate, and Po0.001 for titers) and detection rate (Po0.05) alone were significantly higher in AML and MDS, respectively, than in healthy volunteers (Figure 1b). As for IgG3 WT1 antibody, only patients with CML had significantly higher detection rate (Po0.05) and antibody titers (Po0.05) than healthy volunteers (Figure 1c). Neither detection rate nor antibody titers of IgG4 WT1 antibody were significantly different between the patients and Leukemia
Th1-biased immune responses against WT1 F Wu et al
270 Table 1
Summary of detection rate of IgG subclasses of WT1 antibody in healthy volunteers and patients IgG1
Healthy volunteers
IgG2
4/53 (7.5)
5/53 (9.4)
Patients AML ALL CML CP BC
35/96 12/28 2/12 5/16 1/8 4/8
MDSa IPSS Early stage Advanced stage
14/36 (38.9)*** 4/18 (22.2) 10/18 (55.6)**** 16/40 1/15 8/15 7/10
FAB RA RAEB RAEB-t
IgG3
(36.5)**** (42.9)*** (16.7) (31.2)** (12.5) (50.0)***
23/96 9/28 1/12 3/16 1/8 2/8
5/53 (9.4)
(24.0)* (32.1)* (8.3) (18.8) (12.5) (25.0)
19/96 5/28 1/12 5/16 1/8 4/8
9/36 (25.0)* 4/18 (22.2) 5/18 (27.8)
(40.0)*** (6.7) (53.3)**** (70.0)****
10/40 4/15 3/15 3/10
IgG4
(25.0)* (26.7) (20.0) (30.0)
8/53 (15.1)
(19.8) (18.9) (8.3) (31.3)* (12.5) (50.0)**
24/96 9/28 3/12 4/16 1/8 2/8
(25.0) (32.1) (25) (25.0) (12.5) (25.0)
8/36 (22.2) 3/18 (16.7) 5/18 (27.8)
8/36 (22.2) 4/18 (22.2) 4/18 (22.2)
8/40 2/15 3/15 3/10
8/40 5/15 1/15 2/10
(20.0) (13.3) (20.0) (30.0)
(20.0) (33.3) (6.7) (20)
Numbers in the parenthesis are percentages of detection rate. P-values shown in the table are those for difference in detection rates of WT1 antibody between the individual disease and healthy volunteers. *Po0.05; **Po0.01; ***Po0.001; ****Po0.0001. The detection rates significantly higher than those in healthy volunteers are shown in bold. aThe MDS patients were fractionated into early (Low risk and Int-1) and advanced (Int-2 and High risk) stages according to IPSS (International Prognostic Scoring System), or RA, RAEB, and RAEB-t according to FAB (French–American–British) classification. Four of 40 patients could not be classified by IPSS because of the lack of some information needed for it.
a
**** p< 0.001 *** p< 0.01 **
85.0
p< 0.001 ***
70.0 55.0 40.0
p< 0.05
*** n.s.
120.0
n.s. p
